Intentional illiquidity, unintended costs

Token locking is everywhere in DeFi — vote-escrow systems, staking contracts, vesting schedules. The illiquidity is the point: capital gets locked for a duration to enforce commitment and prevent short-term speculation. That's the design.

The cost falls unevenly. Holders who need to exit before maturity have no satisfactory option. They can wait, or they can resort to OTC deals, bespoke wrappers, and off-chain arrangements that offer thin liquidity, poor price discovery, and no link back to the mechanism that enforces the lock. The result is a persistent tension between incentive alignment and capital efficiency.

Trading the claim, not the asset

Harbor rests on a single observation: a time-locked position can be made liquid without weakening the lock, as long as what trades is the claim on future redemption — not the underlying asset itself. The issuer contract still governs when the underlying can be redeemed. Only the identity of the eventual recipient changes.

The mechanism: an escrow contract absorbs heterogeneous bond NFTs — locked positions with face values and maturities — and mints a single fungible ERC-20 claim token, $\hat{X}$, backed one-to-one by the underlying collateral:

$$X \leftrightarrow \hat{X}$$

A market contract prices bond NFTs against $\hat{X}$ via configurable bid/ask curves. Because $\hat{X}$ is a standard ERC-20, it can be listed as an $X / \hat{X}$ pair on any existing DEX — Uniswap, PancakeSwap, CoW Protocol — and inherit the deep liquidity those venues already provide.

Self-extinguishing by design

Every bond in the escrow eventually matures. As bonds settle, locked collateral becomes liquid $X$ available for 1:1 exchange with $\hat{X}$; the claim tokens are burned. The system's obligations are self-extinguishing — permanent insolvency is impossible by construction. Under stress, $\hat{X}$ may trade below 1:1 with $X$ temporarily, but the discount is bounded by the time remaining until the latest-maturing bond.

The market quotes bid and ask prices $P_b(t)$ and $P_s(t)$ against each bond's remaining value $R$, subject to three invariants:

$$P_b(t) \leq P_s(t) \leq R, \quad \frac{dP(t)}{dt} \geq 0$$

Spread first, boundedness second, monotonicity third: prices don't cross, never exceed guaranteed redemption, and converge upward toward par as maturity approaches.

A new information signal

Just as spreads on corporate debt encode market beliefs about issuer creditworthiness, the discount at which X̂ trades against X — and the prices at which bond NFTs clear — encode participant confidence in the underlying protocol. A token's spot price already reflects aggregate sentiment; it doesn't isolate the cost of illiquidity from beliefs about the protocol itself. Harbor surfaces that second component directly, giving issuers feedback that was previously unavailable.

Status

Harbor is in active development. Documentation is live, and the implementation is open-source on GitHub. The mechanism is described in full in the Trading Time litepaper. For DAOs and protocols considering Harbor, get in touch.